Pupils vs machine
5#Pupils vs Machine
5.2 Lesson planning
Since the modules do not follow on from each other, the #Pupils vs Machine module can be used without any of the other modules having been covered.
This module has been inspired by ‘Le jeu de Nim pour comprendre l’apprentissage par renforcement’ of La Scientotèque in Brussels and by ‘Quand une machine apprend à jouer au jeu de Nim’ of Marie Duflot-Kremer (University of Lorraine).
The game is also mentioned in the updated Media Compass as inspiration for lesson planning.
- Target group: 7e–5e (first to third year of secondary education)
- Room: normal school room
- Equipment per group:
-
8 cups (numbered from 1 to 8)
-
8 pens
-
8 tokens or balls with the number ‘1’
-
7 tokens or balls with the number ‘2’
-
6 tokens or balls with the number ‘3’
-
- Time: one lesson
Artificial intelligence has infiltrated many areas of our lives – from self-driving cars, to robots, through to streaming platforms that recommend new films, series or music (du Sautoy, 2019). Even though it is all around us, artificial intelligence still often scares us: We tend to personalise machines and robots and feel less intelligent than computers (Alexandre et al., 2021). A survey by the Luxembourg Institute of Socio-Economic Research at the University of Luxembourg found a large proportion of people in Luxembourg think AI will make everyday tasks easier for us (70%) and eliminate repetitive tasks at work (64%) – but 64% also think that AI can’t be trusted. More than half consider AI use not to be transparent, and they don’t know exactly where and when AI is even being applied (Poussing, 2021).
Not all of these fears are completely irrational. In her book ‘Weapons of Math Destruction’, mathematician Cathy O’Neil describes various examples of negative consequences of AI algorithms (O’Neil, 2016). The article ‘Why it is best not to let the computer decide’ also describes potential risks and dangers of AI algorithms, such as those that automatically grade pupils’ homework without a proper understanding of what the grade is based on.
However, to separate real dangers from irrational fears, an enlightened approach through critical understanding is needed, as discussed by Alexandre et al. (2021). Jean-Gabriel Ganascia puts it like this: ‘The world is changing and we can’t get stuck in our ways. But we shouldn’t be afraid either. Although the risks are undeniable, there are also some very positive aspects. We will get the most out of technology if we can anticipate the risks.’¹
It is therefore important to start instilling these beliefs at an early age. This module is not intended as a substitute for a course on AI, nor does it teach the technical or mathematical aspects of AI. Rather, it gives the pupils a feel for how it works. The activity is designed to show them that, although an AI can find a winning strategy for a game, this does not make it directly superior to us humans.
1 Faut-il redouter l‘intelligence artificielle? https://uclouvain.be/fr/decouvrir/louvains/faut-il-redouter-l-intelligence-artificielle.html
The pupils understand and reflect on how an AI system processes information to develop a winning strategy.
The pupils define ‘intelligence’ and understand the difference between human and artificial intelligence.
Learning objectives from the Media Compass¹
- Competences 1 – Information and data: 1.2 Analysing and assessing data, information and digital content
- Competences 2 – Communication and collaboration: 2.1 Working with others
As shown in the contextualisation of knowledge, the topic of artificial intelligence is omnipresent. This makes it all the more important to educate people on AI, and to teach a basic understanding of its background and how it works.
The aim of this module is not to understand the technical and/or mathematical aspects of AI, but rather, for the pupils to become aware of how an AI system ‘draws conclusions’ and ‘makes decisions’. It is about depersonalising AI and showing the pupils that an AI system follows an algorithm and does not draw its own conclusions. This is the only way to stop the pupils seeing AI as a black box that works in ways that are impossible to understand.
Because the pupils play the role of the AI themselves during the game and merely follow the algorithm, they can understand the concept more easily. To highlight this specific aspect, the module has also been designed as an analogue module, i.e. a ‘computer science unplugged’ activity. In this way, the pupils are not distracted by screens, codes and other things and can concentrate fully on how AI works.
The module ends with a discussion on the difference between human and artificial intelligence. In some respects, ‘artificial intelligence’ is a poor choice of words, as many people immediately imagine it to mean a superior intelligent robot like the ones in science fiction films. It is therefore important that the pupils know at the end of the module what artificial intelligence means and how it (still?) differs from human intelligence.
Explanation of the rules
First, the game Nim is introduced. The rules are then explained: These can be found under 5.3 M1.
There are many variants of Nim: the number of pens can vary, the player who is last to take away one or more pens wins or loses. The game can be made slightly more complicated by arranging the pens in a pyramid shape and only being allowed to take away pens from the same row at a time. For this module, we have chosen the simpler variant with a small number of pens.
The pupils split into groups of two and play a few rounds to become familiar with the game.
Pupils against machine
In the second phase, the class is divided into groups of 4 pupils. In each group, one pupil plays the role of the machine. The rules are now as follows:
The pupils play together against the machine and try to beat it. The pupils always go first and follow the rules of the game. The machine follows the instructions described in 5.3 M2. The game is played until the machine only has one go left. The machine should then have reached the following position:
| Cup | Tokens |
|---|---|
| 1 | 1 |
| 2 | 2 |
| 3 | 3 |
| 4 | turned over |
| 5 | 1 |
| 6 | 2 |
| 7 | 3 |
| 8 | 1, 2, 3 |
The different processes are summarised again in the following videos (in English and French) from La Scientothèque in Brussels.
The first 20 minutes of the following presentation by the lecturer Marie Duflot-Kremer explain the game in more detail.
Alternative :
You can also let the machine go first. It will give up after a while, when it realises there is no winning strategy. The first cup (i.e. cup number 8) will eventually be turned over.
More difficult alternative:
You can also play the game with more pens. As long as you take a multiple of 4 and let the machine go second, you will get the same result. It will just take longer.
Is the machine superior to us?
The pupils are split into new groups and asked to discuss how the machine came up with a winning strategy so quickly. They should answer the following questions:
- Does the machine even know the rules of the game?
- If the exact same machine plays another game, will it also win?
- Can the exact same machine take your maths test next week?
Answers to these questions must be written down by each group together with explanations.
In the subsequent summary in the class group, it is explained that the machine wins because it has been programmed for the game Nim. However, the same machine would not necessarily be able to win at a different game or take our maths test.
Further resources and workshops
- La Scientothèque in Brussels provides a range of resources and information about artificial intelligence, from definitions through to its history and uses.
- The University of Luxembourg’s Scienteens Lab also offers workshops. The workshop ‘Art et Intelligence Artificielle’ starts with an introduction to the topic of artificial intelligence, before covering the art part in the next phase. The workshop is an excellent follow-up to this module, and the first phase can be enhanced and covered in greater detail if the pupils have already taken this module.
- The website ‘I Am A.I.’ contains numerous learning resources on the topic of artificial intelligence. The project ‘I AM A.I.’ is one of the many initiatives of the non-profit organisation Imaginary, which is dedicated to teaching modern mathematics. On the ‘I AM A.I.’ website, there is a virtual interactive tour, with three different interactive activities. The team has also created a travelling exhibition on the theme ‘I AM A.I.’. Why not check if and when this exhibition is in your area, or even contact ‘I AM A.I.’ and have some exhibits brought to your school?
As indicated above, the game Nim can be made more complex by increasing the number of pens. Groups with more able pupils can also be told to alternate between letting the machine or the pupils go first. This makes it more difficult for the pupils to predict the end.
- The Luxembourgish context: The topic of artificial intelligence and machine learning should not be missing from the school curriculum in the 21st century (see also the Media Compass 2022 update).
- Differentiation: Different variants of the game Nim are described to make it harder and less predictable.
- Media Compass: See the learning objectives of the media literacy framework in the didactic analysis of this document.
- The 4Cs model: Communication, collaboration, creativity, critical thinking. The 4Cs model is taken into account in various ways through the different social forms and teaching activities.
- Relation to current research: The section ‘More on the topic’ provides a link to current research in the field of AI and machine learning.
- Relation to research in Luxembourg: In an interview, a scientist from Luxembourg explains how he uses artificial intelligence and machine learning in his research.
Click on the image to enlarge
References:
Alexandre, Frédéric, Becker, Jade, Comte, Marie-Hélène, Lagarrigue, Aurélie, Liblau, Romain, Romero, Margarida & Viéville, Thierry. (2021). Why, What and How to help each Citizen to Understand Artificial Intelligence? KI – Künstliche Intelligenz, 2, 1610–1987.
O’Neil, Cathy. (2016). Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy. New York: Crown.
Poussing, Nicolas. (2021). Résultats de la consultation publique relative aux opportunités et aux défis de l’Intelligence Artificielle (IA). LISER. https://gouvernement.lu/dam-assets/documents/actualites/2021/04-avril/28-bettel-cdp/Rapport-IA6-final.pdf
du Sautoy, Marcus. (2019). The Creativity Code: Art and Innovation in the Age of AI. Cambridge: Belknap of Harvard UP.
